Extraction of fuzzy object information in multidimensional images for quantifying MS lesions of the brain
A technique for object information extraction from images which retains fuzziness as realistically as possible. The technique is used for image segmentation of fuzzy objects for n-dimensional digital spaces based on the notion of "hanging togetherness" of image elements specified by their fuzzy connectedness. A specified fuzzy object is extracted and all fuzzy objects present in the image data are identified by segmenting the image using the notion of "hanging togetherness" of image elements specified by their fuzzy connectedness as defined herein. The technique is used in a preferred embodiment to quantify MS lesions of the brain via magnetic resonance imaging.
Claims
1. An imaging system for identifying fuzzy or blurred objects within a multidimensional scene, comprising:
- means for creating a digital representation of said scene, said digital representation comprising a plurality of spatial elements;
- a memory for storing said digital representation of said scene;
- means for determining a strength of connectedness of each spatial element in said digital representation of said scene with other spatial elements in said digital representation of said scene, where the strength of connectedness along a path between a first spatial element and a second spatial element is defined as the smallest affinity between respective spatial elements along said path;
- means for clustering those spatial elements in said memory having strengths of connectedness with other spatial elements in said memory above a predetermined threshold into a fuzzy connected component of a fuzzy object in said scene; and
- means for displaying said fuzzy connected component so that said fuzzy connected component is distinguished from other fuzzy connected components in said scene.
2. An imaging system as in claim 1, further comprising means for determining a volume distribution of said fuzzy connected component for said predetermined threshold.
3. An imaging system as in claim 1, wherein said means for determining the strength of connectedness of spatial elements in said digital representation of said scene determines the strength of connectedness between said first spatial element and said second spatial element as the largest of the strengths of connectedness of respective paths from said first spatial element to said second spatial element.
4. An imaging system as in claim 1, wherein said means for creating a digital representation of said scene creates said scene as a set of image slices.
5. A method for identifying fuzzy or blurred objects within a multidimensional scene, comprising the steps of:
- scanning said scene to create a digital representation of said scene, said digital representation comprising a plurality of spatial elements;
- storing said digital representation of said scene;
- determining a strength of connectedness of each spatial element in said digital representation of said scene with other spatial elements in said digital representation of said scene, where the strength of connectedness along a path between a first spatial element and a second spatial element is defined as the smallest affinity between respective spatial elements along said path;
- creating a K.sub.o scene comprising spatial elements having strengths of connectedness with other spatial elements above a predetermined threshold; and
- displaying said K.sub.o scene.
6. A method as in claim 5, comprising the further step of determining a volume distribution of a fuzzy connected object comprising spatial elements with strengths of connectedness with other spatial elements above said predetermined threshold.
7. A method as in claim 5, wherein said strength of connectedness determining step comprises the step of determining the strength of connectedness between said first spatial element and said second spatial element as the largest of the strengths of connectedness of respective paths from said first spatial element to said second spatial element.
8. A method as in claim 5, wherein said scanning step comprises the step of creating a set of image slices.
9. A system for detecting multiple sclerosis (MS) lesions in a magnetic resonance (MR) image of a patient's brain, comprising:
- an MR scanner for creating a digital representation of an image slice through the patient's brain, said digital representation comprising a plurality of spatial elements;
- a memory for storing said digital representation of said image slice;
- processing means for determining a strength of connectedness of each spatial element in said digital representation of said image slice with other spatial elements in said digital representation of said image slice where the strength of connectedness along a path between a first spatial element and a second spatial element is defined as the smallest affinity between respective spatial elements along said path, and for clustering those spatial elements having strengths of connectedness with other spatial elements above a predetermined threshold into a fuzzy connected component of the portion of the patient's brain in said image slice; and
- means for permitting identification of said fuzzy connected component as an MS lesion by displaying said fuzzy connected component so that it is distinguished from other fuzzy connected components in said image slice of the patient's brain.
10. A system as in claim 9, further comprising means for determining a volume distribution of said fuzzy connected component for said predetermined threshold, thereby determining a volume distribution of said MS lesion.
11. A system as in claim 9, wherein said processing means determines the strength of connectedness between said first spatial element and said second spatial element as the largest of the strengths of connectedness of respective paths from said first spatial element to said second spatial element and clusters those spatial elements having strengths of connectedness with other spatial elements above said predetermined threshold into said fuzzy connected component of the portion of the patient's brain in said image slice.
12. A system as in claim 9, wherein said MR scanner creates said scene as a set of image slices and said processing means determines a strength of connectedness of each spatial element in a digital representation of one image slice with other spatial elements in a digital representation of another image slice.
13. A method of detecting multiple sclerosis (MS) lesions in a magnetic resonance (MR) image of a patient's brain, comprising the steps of:
- creating a digital representation of an MR image slice through the patient's brain, said digital representation comprising a plurality of spatial elements;
- storing said digital representation of said MR image slice;
- specifying spatial elements in at least one of the gray matter, white matter, and ventricles of the MR image slice through the patient's brain as starting points for a strength of connectedness determination for said spatial elements;
- determining a strength of connectedness of each spatial element in said digital representation of said MR image slice, starting with said starting points, with other spatial elements in said digital representation of said MR image slice, where the strength of connectedness along a path between a first spatial element and a second spatial element is defined as the smallest affinity between respective spatial elements along said path;
- clustering spatial elements having strengths of connectedness with other spatial elements above predetermined thresholds into fuzzy connected white matter, gray matter, ventricle, and lesion components of the portion of the patient's brain in said image slice; and
- identifying said fuzzy connected lesion component as an MS lesion by displaying said fuzzy connected lesion component so that it is distinguished from said fuzzy connected white matter, gray matter, and ventricle components in said image slice of the patient's brain.
14. A method as in claim 13, wherein said clustering step comprises the steps of dividing said digital representation of said MR image slice into segments, computing the center of mass and principal axes of one of said segments, and determining from said principal axes and said center of mass a cluster representing white matter and a cluster representing gray matter of the patient's brain.
15. A method as in claim 14, wherein said clustering step comprises the further step of determining a cluster representing cerebrospinal fluid of the patient's brain by computing the center of mass of another one of said segments predicted to include said cerebrospinal fluid and determining which fuzzy connected component contains said center of mass of said another segment.
16. A method as in claim 13, comprising the further step of determining a volume distribution of said fuzzy connected lesion component for a predetermined threshold, thereby determining a volume distribution of said MS lesion.
17. A method as in claim 13, wherein said strength of connectedness determining step comprises the steps of determining the strength of connectedness between said first spatial element and said second spatial element as the largest of the strengths of connectedness of respective paths from said first spatial element to said second spatial element and clustering those spatial elements having strengths of connectedness with other spatial elements above said predetermined threshold into said fuzzy connected lesion component of the portion of the patient's brain in said image slice.
18. A method as in claim 13, wherein said digital representation creating step comprises the step of creating said scene as a set of image slices and said strength of connectedness determining step comprises the step of determining a strength of connectedness of each spatial element in a digital representation of one image slice with other spatial elements in a digital representation of another image slice.
| 4545384 | October 8, 1985 | Kawachi |
| 4941190 | July 10, 1990 | Joyce |
| 4991092 | February 5, 1991 | Greensite |
| 4991224 | February 5, 1991 | Takahashi et al. |
| 5018214 | May 21, 1991 | Pasch |
| 5390258 | February 14, 1995 | Levin |
- Udupa and Herman, 3D Imaging in Medicine, Boca Raton, FL: CRC Press, 1991. Hohne, K. et al., 3D Imaging in Medicine, Algorithms, Systems, Applications. Berlin: Springer-Verlag, 1990. Kaufman, A., A Tutorial on Volume Visualization. Los Alamitos, CA: IEEE Computer Society Press, 1990. Levoy, M., "Display of surfaces from volume data," IEEE Computer Graphics and Applications, vol. 8(3), pp. 29-37, May 1988. Drebin, R. et al., "Volume rendering," Computer Graphics, vol. 22, pp. 65-74, 1988. Vannier, M. et al., "Multispectral magnetic resonance image analysis," CRC Critical reviews in Biomedical Engineering, vol. 15, pp. 117-144, 1987. Udupa, J. et al., "Surface and volume rendering in 3D imaging: a comparison," Journal of Digital Imaging, vol. 4, pp. 159-168, 1991. Udupa and Odhner, "Shell rendering," IEEE Computer Graphics and Applications, vol. 13, No. 6, pp. 58-67, 1993. Rosenfeld, A., "Fuzzy digital topology," Information and Control, vol. 40, No. 1, pp. 76-87, 1979. Rosenfeld, A., The fuzzy geometry of image subsets, Pattern Recognition Letters, vol. 2, No. 5, pp. 311-317, Sep., 1984. Reder and Arnason, Immunology of Multiple Sclerosis. In: Vinken, P.J. et al., Handbook of Clinical Neurology, vol. 3 (47): Demyelinating Diseases. Amsterdam: Elsevier Science, pp. 337-395, 1985. Young, I. et al., "Nuclear magnetic resonance imaging of the brain in multiple sclerosis," Lancet, vol. 2, pp. 1063-1066, 1981. Willoughby, E. et al., "Serial magnetic resonance scanning in multiple sclerosis: A second prospective study in relapsing patients," Ann. Neurol., col. 14, pp. 43-49, 1989. Wolff and Balaban, "Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo," Magn. Reson. Med., vol. 10, pp. 135-144, 1989. Dousset, V. et al., "Experimental allergic encephalomyelitis and multiple sclerosis: Lesion characterization with magnetization transfer imaging," Radiology, vol. 182, No. 2, pp. 483-491, 1992. Outwater, E. et al., "Magnetization transfer of hepatic lesions: Evaluation of a novel contrast technique in the abdomen," Radiology, vol. 182, pp. 535-540, 1992 McGowan, J. et al., "Magnetization transfer imaging with pulsed off-resonance saturation: Contrast variation with saturation duty cycle," J. Mag. Res., vol. 4, No. 1, pp. 79-82, 1994. Kapouleas, I., "Automatic detection of white matter lesions in magnetic resonance brain images," Computer Programs and Methods in Biomedicine, vol. 1, pp. 17-35, 1990. Raya, S., "Low-level segmentation of 3-D magnetic brain images-a rule-based system," IEEE Transactions on Medical Imaging, vol. 9, No. 3, pp. 327-337, 1990. Kohn, M. et al., "Analysis of brain and cerebrospinal fluid volumes with MR imaging. Part I. Methods, reliability and validation," Radiology, vol. 178, pp. 115-122, 1991. Kamber, M. et al., "Model-based 3-D segmentation of multiple sclerosis lesions in dual-echo MRI data," Proceedings of VBC'92, pp. 590-600, 1992. Bezdek, J. et al., "Review of MR image segmentation techniques using pattern recognition," Medical Physics, vol. 20, No. 4, pp. 1033-1048, 1993. Soltanian-Zadeh, H. et al., "A comparative analysis of several transformations for enhancement and segmentation of magnetic resonance image scene sequences," IEEE Transactions of Medical Imaging, vol. 11, No. 3, pp. 302-318, 1992. Cline, H. et al., "Three-dimensional segmentation of MR images of the head using probability and connectivity," JCAT, vol. 14, No. 6, pp. 1037-1045, 1990. Kaufmann, A., Introduction to the Theory of Fuzzy Subsets, vol. I. New York, NY: Academic Press, Inc., 1975. Kong and Rosenfeld, "Digital topology: Introduction and survey," Computer Vision, Graphics, and Image Processing, vol. 48, pp. 357-393, 1989. Herman, G., "Discrete multidimesional Jordan surfaces," CVGIP: Graphic Models and Image Processing, vol. 54, pp. 507-515, 1992. Udupa, J.K., "Multidimensional digital boundaries," CVGIP: Graphical Models and Image Processing, vol. 50, No. 4, pp. 311-323, 1994. Udupa and Ajjanagadde, "Boundary and object labelling in three-dimesional images," Computer Vision, Graphics, and Image Processing, vol. 51, pp. 355-369, 1990. Napel, S. et al., "Visualizing three-dimensional flow with simulated streamlines and three-dimensional phase-contrast MR imaging," Journal of Magnetic Resonance Imaging,vol. 2, pp. 143-153, 1992. Udupa, J. et al., "3DVIEWNIX: An open, transportable, multidimensional, multimodality, multiparametric imaging software system," in SPIE Proceedings, vol. 2164, pp. 58-73, 1994.
Type: Grant
Filed: Feb 24, 1995
Date of Patent: Sep 22, 1998
Inventors: Jayaram K. Udupa (Audubon, PA), Supun Samarasekera (Philadelphia, PA)
Primary Examiner: Christopher Kelley
Law Firm: Woodcock Washburn Kurtz Mackiewicz & Norris LLP
Application Number: 8/394,231
International Classification: G06K 946;
